Please feel free to respond to as many questions as you have time for.
1. I'm a huge fan of your ML work while you were at Stanford. If I may steal a minute or two from you, I'd like to ask, why do you think recursive neural networks have lost out to other architectures (basically transformers) when it comes to NLP, where intuition would suggest the composionality of language would favor recursive NNs? Do you think recursive NNs will make a come-back sometime?
2. Also, you say on HN that you.com is "the world's first open search engine platform that summarizes the web for you". Do you mean this in the sense of summarization in the NLP sense, or some other interpretation? I did a couple of searches on you.com, and I didn't notice anything that "summarized" (in the NLP sense).
3. From the business side, what's to stop Google from straight up copying your features if you gain traction? They can certainly afford to pay for talent, and are not known to be sloths when it comes to new tech.